Chapter 5. Running Dakota

How to use run configurations

To run Dakota using the GUI, you will need to set up a run configuration.  Fortunately, this is a somewhat straightforward process for most cases, and is designed to save you lots of time in the future.

The following example screenshots use the example input file, available in the Dakota examples repo (see Chapter 3).

To begin, select Run > Run Configurations from the dropdown menu...


...or select the small down arrow next to the Run icon and select Run Configurations.


(If these options are not available, turn them on by going to Window > Perspective > Customize Perspective.  Then, select the Action Set Availability tab and make sure “Launch” is checked underneath the “Available action sets” column.)

The Run Configurations dialog allows you to specify a template for running Dakota.  This only needs to be done once; after the first time, all your run configuration information is preserved, making a run of Dakota as easy as pushing one button.  You can make as many run configurations as you want.


To begin, double-click on the Dakota line in the list on the left.

Dakota Run Configuration options


Name:  The name of your run configuration.  This can be whatever you want.

Main Tab:

  • Dakota Settings:
    • Run Directory:  The directory in which Dakota will run.  This can be anywhere on your filesystem.
    • Input File:  The Dakota input file to run.  This field is usually auto-populated for you (see the section on Run Shortcut Locations below) but if it's not, use the file browser button to get the input file you want to run.
    • Program Arguments:  Any additional command-line arguments you would like to pass to Dakota.  If you run “dakota -help” on the command line outside of the GUI, Dakota will give you a list of all possible command line arguments.
  • Python Settings:
    • Run Directory:  The directory in which Python will run, if you have a Python interface script specified for your Dakota study.  Specifying this path is required to get a Python Dakota interface script to run (see Chapter 6).  This field essentially has the same effect as specifying a PYTHONPATH environment variable outside of the Dakota GUI.
      • Note:  If you are using the "" example input file shown in the screenshots, this field can be left blank.


Import Tab:

When Dakota is finished running, the GUI can automatically import any files that Dakota has generated and store them in a timestamped directory for future reference.  This is a handy feature if you want to preserve information about every run you do with Dakota, because on subsequent runs of Dakota, Dakota will automatically write over any previous files in its specified run directory.

  • Parent directory:  The name of the directory to back up files to.  The default name is “run_results” but you can choose any name you like.
  • Import checkboxes:  Use these checkboxes to specify which of Dakota’s generated files you want to make backup copies of.  If you don’t want the GUI to back anything up, you can uncheck all of these boxes.

When you are done configuring your Dakota run configuration, click the Apply button to save your changes.

Note:  In all cases for all run configurations, the Dakota GUI will use the globally specified version of Dakota as the target process to run.  This can be configured in Windows > Preferences > Dakota (see Chapter 2).

When you are ready to run, click Run.

After the First Run

After configuring this run configuration once, you can access it easily by clicking the down arrow near the Run icon:


This menu keeps a history of recent previous run configurations.  Simply select the one you want to run – there is no need to open the Run Configurations dialog again unless you want to reconfigure your run configuration.

Dakota Console options

Dakota displays its console output in the Console view:


There are some actions available on the Console view:


From left to right:

  • Terminate:  Attempts to stop Dakota from running further.  This icon will be grayed out if Dakota is not running.
  • Remove Console:  Close the console view.  You can always bring the console back from the list of views (Window > Show View > Other… then select General > Console).
  • Clear Console:  Rather than closing the whole console, this action simply clears all of the text and leaves the console open.
  • Scroll Lock:  Toggle scroll lock on the console.
  • Word Wrap:  Toggle word wrap for the console.
  • Pin Console:  Pins the console to the front.  If you have multiple consoles open, pinning keeps the current one in focus.
  • Display Selected Console:  Allows you to switch between multiple open consoles.
  • New Console View:  Opens a new console.

Run shortcut locations

You may not care to mess with run configurations for running Dakota.  For the most basic Dakota input file examples that don’t require any additional run configuration setup, there are some shortcut options.

  • Right-click on your Dakota study, then select Run As > Dakota.
  • Or, you can right-click on the Dakota study itself and choose Run As > Dakota.

Doing this will automatically create a new run configuration for you with the following configuration:

  • The Dakota input file will be the one you selected.
  • Dakota’s run directory will be the parent directory of the input file.
  • All possible output files will be imported.

Note:  This run configuration doesn’t specify a Python run directory.  If you use this option for a more complex Dakota input file that uses a Python driver (for instance, the cantilever beam test problem in Chapter 6) this run configuration shortcut will not work and will require manual configuration from the user.